The importance of the novel heteromeric, basolateral transporter, Ostα-Ostβ, in enterohepatic circulation of bile acids and the homeostasis of bile acid synthesis has recently been confirmed . Although it is clear that function of this facilitated transporter requires expression of both subunits, it is not known whether functional activity depends upon (1) the acquisition of N-glycosylation of the alpha subunit, (2) the beta subunit for its ability to release the alpha subunit from an ER retention signal, or (3) the physical interaction of the two proteins at the plasma membrane. The data provided here indicate that glycosylation of OSTα is not necessary for transporter localization or function. Furthermore, it shows that the physical interaction of the two subunits may be transient, suggesting that association at the plasma membrane may not be necessary for transporter function.
Glycosylation of a protein is one of the major biosynthetic functions of the ER and is a common post-translational modification of membrane proteins. Although the addition of the "core" oligosaccharide occurs in the ER, further extensive processing or trimming occurs in the Golgi and results in what is commonly referred to as the complex or mature glycoprotein . N-glycosylation is found usually in the sequences Asn-X-Ser or Asn-X-Thr, where X is any amino acid [13, 14]. Although this consensus motif is found in the N-terminus of the alpha subunit in the mouse, rat and skate, it is not present in the human OSTα . Instead, the sole asparagine residue in an extracellular site is in the sequence Asn25-X-Gly in the N-terminus. We have shown in this study that, despite the lack of traditional consensus sequence, human OSTα is expressed on the cell surface as a glycoprotein. Similar to previous reports [3, 4, 6, 7, 15] our data indicate that endogenous alpha subunit migrates in SDS-PAGE as a single band and precursor forms are not detected. This suggests that in the presence of the beta subunit the glycoprotein is efficiently trafficked through the Golgi. It is only in the over-expressing transfected cells that the multiple forms of the alpha subunit are seen (Figure 3, 4 and 5 this manuscript; [3, 6].
The necessity for glycosylation of proteins has been studied for many years and is largely believed to be important in proper folding and stabilization of newly synthesized proteins and in affecting the charge and solubility of the protein [16, 17]. The critical nature of this folding is highlighted by the finding that detection of misfolded glycoproteins in the ER can result in ER-associated degradation (ERAD) [18, 19]. Our data indicate that the lack of oligosaccharide chain on the alpha subunit does not designate the polypeptide for ERAD. Instead, after tunicamycin inhibition of glycosylation, the transporter was still trafficked properly to the plasma membrane where it was fully functional, indicating that interaction between the alpha and beta subunits is not compromised by the lack of oligosaccharide. Perhaps because the alpha subunit of the organic solute transporter has only one asparagine residue in an extracellular domain, the affect of the absence of the carbohydrate on folding is not critical. Tunicamycin treatment has been used to study glycosylation of other hepatocyte proteins. The absence of oligosaccharide did not affect the secretion of transferrin or very low density lipoprotein, but did interfere with the ability of the apical membrane protein, Mrp2, to be trafficked to the plasma membrane in rat hepatocytes . And recently the N-linked carbohydrates have been described for the hepatocyte basolateral membrane protein oatp1a1 and found to be important in the protein's localization and function . In HepG2 cells it has been reported that five of eight glycoproteins studied did not require glycosylation for their trafficking . Mochizuki et al have shown that rat Bsep requires at least two of its four N-linked glycans for proper protein stability, intracellular trafficking and functional activity .
Interestingly, we (Figure 5B) and others [5, 7] have shown that the absence of one of the subunits leads to degradation of the other subunit. Thus, it is the presence and interaction of the two subunits that are critical to the stability of the heteromeric, intact transporter, and not the glycosylation of the alpha subunit. Protein-protein interactions in the ER are known to be critical for many different processes, including trafficking and function of multimeric membrane proteins. The presence of fully functional oligomeric complexes at the plasma membrane can involve specific ER retention/retrieval motifs[24, 25], anterograde ER export signals [26, 27], interaction with scaffold protein [28–31], and phosphoylation [29, 32]. The necessity for interaction between OSTα and OSTβ subunits in the ER suggests that physical association of the two proteins may mask a retention/retrieval motif or, alternatively, may reveal a forward trafficking motif. The RXR motif is one such retention/retrieval sequence and it is interesting that both the alpha and beta subunits contain an RXR-like motif in their C-terminal sequence. It remains to be determined whether this sequence is important in the localization of the organic solute transporter.
Our immunoprecipitation data confirm that the OSTα and OSTβ interaction is essential early in the biosynthetic process, but suggest that it may not be necessary later once the major protein gets to the plasma membrane. Because the only way to get OSTα to the plasma membrane is to co-express the beta subunit, it is impossible to determine if the alpha subunit actually requires the beta subunit for its functional activity. However, the lack of co-precipitation between the mature form of OSTα and the OSTβ subunit suggests that this may not be the case. When Li and colleagues  performed similar immunoprecipitations in HEK293 cells transfected with mouse Ostα and Ostβ constructs, they also saw only a single band after precipitation with anti-Myc. However, they indicate that it is the mature form of the protein. Given that all data point to the interaction of the subunits in the ER, one would also expect to see the immature form precipitated. Similarly, in mouse ileum Li et al show only one band for Ostα on Western blots and this protein is co-precipitated by an antibody to Ostβ . Although the explanation for these differences in immunoprecipitation is still unclear, we cannot discount that it is due to species variability or species-specific antibody affinity.
The possible transient nature of the subunit interaction also appears to be in conflict with immunofluorescent studies which suggest co-localization of the subunits at the plasma membrane in transfected cells (Figure 2 and . However, the finding of a yellow color indicating co-localization may be due to the close proximity of the two subunits, not the actual association. Optical microscopes are unable to resolve two items that are closer together than 200 nm. Also, we cannot discount the possibility that, similar to tunicamycin treated cells, some "immature" protein might be expressed on the plasma membrane, and, thus, be detected by the primary antibodies. Bimolecular fluorescence complementation has also been used to study the interaction of the two subunits in HEK293 cells transfected with mouse Ostα and Ostβ . These studies clearly show that complementation occurs between Ostα and Ostβ and results in plasma membrane localization. However, the possibility that the interaction might be transient cannot be assessed because, once the complementation reaction occurs, it is irreversible.